LetsLaunch

How to Launch a Two-Sided Marketplace: Solving the Chicken-and-Egg Problem (Notes From Launching One)

Pick the side that is harder to get, give it a reason to show up before the other side exists, and bring in the first demand by hand. The four decisions behind launching a services marketplace, and what to measure in the first 90 days.

Costin Gheorghe10 min read

How to launch a two-sided marketplace, in one sentence: pick the side that is harder to get (usually supply), give it a reason to show up before demand exists, and bring in the first demand by hand rather than waiting for it. Almost every other piece of advice about the chicken-and-egg problem is a variation on those three moves.

I'm writing this from the middle of doing it. In October 2026 I launched Contractez, a Romanian services marketplace (the site is in Romanian): clients post a project, and compatible providers (freelancers, tradespeople, small firms) send offers. It covers everything from electricians, plumbers and renovations to web development, design, accounting and tutoring. It's early, so there are no growth charts here. What follows are the four decisions I made before launch, the reasoning behind each, and the numbers I'm watching to find out whether they were right.

What is the chicken-and-egg problem in a marketplace?

A marketplace is only worth joining if the other side is already there. Clients won't post projects on a site with no providers, and providers won't build a profile on a site with no projects. Each side waits for the other, so on launch day both sides look at an empty room and leave. A one-sided product can win its first user with a good feature. A marketplace has to win two groups of people at once, and neither group cares how good the software is until the other one shows up.

The problem doesn't end at launch, either. It repeats in every new city and every new category: a marketplace with plenty of plumbers in one city is still empty for a client looking for a translator in another.

How to launch a two-sided marketplace: which side goes first?

Start with the side that is harder to get. James Currier's 19 tactics for the chicken-or-egg problem at NFX lists "Get the hardest side first" as tactic number one, and Lenny Rachitsky, after interviewing the people who built Airbnb, Etsy, Thumbtack and others, writes that "the vast majority of marketplaces start off supply-constrained".

In a services marketplace, the hard side is supply. The asymmetry is easy to see once you picture both failures:

  • A client who posts a project and gets zero offers leaves, and usually doesn't come back. They had a real need (a leaking pipe, a website to finish) and the site failed them on it. Next time they ask a Facebook group or a friend.
  • A provider who joins and sees few projects for a week is disappointed, but the cost of waiting is low. Their profile stays up, and if alerts tell them when something relevant arrives, they don't have to check the site every day.

So my plan for Contractez is supply first: recruit the first providers before pushing for clients, so that the first projects posted actually get offers. Demand comes second, and comes from places where people are already asking for help (more on that below).

One more thing that follows from "hardest side first": the easy side needs to be served well when it does arrive. If you recruit providers for three months and the first ten clients all wait days for an offer, you've burned the demand you worked to attract. Connecting the first clients to providers by hand is tactic 13 on the NFX list, and it's the one founders most often skip because it doesn't scale. It doesn't need to. It needs to make the first transactions happen.

How do you give the supply side a reason to join?

Supply that joins before demand exists needs a reason that works on day one. The classic examples give one side a standalone tool: OpenTable sold restaurants reservation software before the diners arrived. For a services marketplace, I looked at the opposite problem: why providers leave.

On large open marketplaces, a freelancer applies to a project alongside dozens of others. Most of those applications are never read. Applying feels like a lottery, so good providers stop applying, and the ones who stay compete on price.

The product rule I built around that: by default, a project on Contractez accepts at most 5 offers. Once five providers have applied, the project closes to new offers, so each one gets read. A client who wants more choice can raise the limit, up to 30. There's a short explanation of how the offer cap works on the site (in Romanian).

It's a trade-off, and it's worth being honest about it:

LeverWhat it gives supplyWhat it costs
Offer cap (5 by default)Each offer is read; applying is worth the effortClients see fewer options unless they raise the cap
Free profile + 5 free applications a monthNo cost to try the platformFree accounts that never apply still cost something to run
Email alerts for compatible projectsNo need to check the site while demand is thinA bad matching rule sends noise and trains people to ignore alerts
Paid plans for more applicationsA way to do more volume once it worksHas to feel worth it, or it reads as a paywall

The cap is also a message to providers: competing with four other people, not forty, is a reason to sign up early. Whether it works is something I can only measure, not assert, which is why it shows up in the metrics section.

How should a new marketplace make money?

Not by taxing the transactions you are trying to create. Contractez is free to start on both sides: posting a project is free, and providers get a free profile with 5 applications a month. It takes no commission on the work. Client and provider agree on the price and pay each other directly.

Revenue comes from providers who want more volume:

  • Pro, 49 lei a month: 40 applications, plus Auto Bid, which applies to matching projects automatically by rules the provider sets.
  • Business, 149 lei a month: 200 applications.

The reasoning: early on, every completed transaction is evidence for both sides that the marketplace works. A commission on those transactions makes each one slightly less likely to happen, and in services it gives both sides a reason to swap phone numbers and finish the deal elsewhere. A subscription charges the providers who are already getting value, at the point where they want more of it.

The cost of this model is that you never hold the money, so you can't use payments as your trust mechanism the way escrow-based platforms do. That pushes trust into the product instead.

How do you build trust before you have reviews?

A new marketplace has no reviews, and reviews are the main thing a client uses to choose a provider. Faking them is out of the question. Two decisions cover the gap:

  1. Reviews only after a collaboration both sides confirm. A review on Contractez can't be posted by someone who never worked with the provider. This makes reviews slower to accumulate, but each one means something, and it closes the obvious abuse of a fresh platform: providers reviewing themselves.
  2. An "ANAF verified company" badge. ANAF is Romania's tax authority, and companies can get a badge showing they're verified against it. It gives a client a real signal on day one, before a provider has a single review.

A related privacy rule: the client's phone number is not shown publicly. People posting a project for their home shouldn't get cold calls for it.

How do you get your first marketplace users without ads?

Paid acquisition before liquidity buys visits to an empty room. The channels I'm using instead cost time rather than money:

  • Communities where people already ask for services. Romanian Facebook groups are full of posts like "can anyone recommend an electrician in Cluj?". The approach is posting with the admins' permission and writing personal, manual messages. No bots and no copy-paste blasts, which get you removed from the group and burn the channel for good.
  • SEO content that matches how people search for services: price guides ("how much does a bathroom renovation cost") and city × service pages. It takes months to pay off, but each page is a small entry point to one niche in one city, which is the "start narrow" advice applied one page at a time.
  • Launch directories. These reach founders and early adopters, not someone looking for a plumber, so their job is early visibility and the first backlinks for a new domain. The order matters more than the number of directories; our startup launch checklist walks through it, and the list of directories that give dofollow links shows which ones are worth the form. If you want to skip the big boards entirely, the Product Hunt alternatives roundup covers the smaller ones. You can also launch on LetsLaunch for free.

The pattern across all three: go where the demand already exists and bring it back, rather than trying to create it.

What should you measure in the first 90 days?

Not signups. Signups on either side can grow while the marketplace is still failing. These are the metrics I'm tracking, framed as what to watch rather than results:

MetricWhy it matters
Completed provider profiles, per category and citySupply that can't be matched to anything doesn't count
Share of projects with at least one offerThe client's first experience; a project with zero offers is a lost client
Time to first offerThe speed at which a client learns the marketplace works
Offers per project, compared with the capTells you whether the cap is binding or supply is thin
Projects that end in a collaboration both sides confirmThe closest thing to a completed transaction when payment happens off-platform

If I had to pick one, it would be the share of projects that get at least one offer. It's the metric that tells you whether supply-first is working, and it's the one most likely to drift when you start pushing demand.

How long does it take to solve the chicken-and-egg problem?

There's no fixed timeline, because it isn't solved once. It's solved niche by niche: one category in one city at a time. The sign you're getting there is that projects in that niche get offers without you stepping in by hand.

Should you fake supply to launch a marketplace?

No. Fake listings, fake profiles and fake reviews get found out, and when they do they cost the trust a new marketplace has least of. Doing things by hand is different and is fine: personally recruiting providers, or connecting the first clients to providers yourself, is real supply delivered manually.

Is commission or subscription better for a new marketplace?

Early on, a subscription on the supply side usually costs less growth than a commission, because it doesn't tax each new transaction. Commission makes more sense once the marketplace adds enough value per deal, such as payment protection, that people accept the fee rather than routing around it.

Can you launch a marketplace without an ad budget?

Yes. Communities where people already ask for what you offer, SEO pages for specific niches, and launch directories all cost time instead of money. Paying for traffic makes more sense after the marketplace can reliably serve the people that traffic brings in.

ShareXLinkedInHacker NewsRedditon LetsLaunch

What's live on LetsLaunch right now